[ruby/prism] Further optimize ripper translator by not using `delegate`

Using it seems pretty bad for performance:

```rb
require "benchmark/ips"
require "prism"
require "ripper"

codes = Dir["**/*.rb"].map { File.read(it) }

Benchmark.ips do |x|
  x.report("prism") { codes.each { Prism::Translation::Ripper.lex(it) } }
  x.report("ripper") { codes.each { Ripper.lex(it) } }
  x.compare!
end
```

```
# Before
ruby 4.0.0 (2025-12-25 revision https://github.com/ruby/prism/commit/553f1675f3) +PRISM [x86_64-linux]
Warming up --------------------------------------
               prism     1.000 i/100ms
              ripper     1.000 i/100ms
Calculating -------------------------------------
               prism      0.319 (± 0.0%) i/s     (3.14 s/i) -      2.000 in   6.276154s
              ripper      0.647 (± 0.0%) i/s     (1.54 s/i) -      4.000 in   6.182662s

Comparison:
              ripper:        0.6 i/s
               prism:        0.3 i/s - 2.03x  slower
# After
ruby 4.0.0 (2025-12-25 revision https://github.com/ruby/prism/commit/553f1675f3) +PRISM [x86_64-linux]
Warming up --------------------------------------
               prism     1.000 i/100ms
              ripper     1.000 i/100ms
Calculating -------------------------------------
               prism      0.482 (± 0.0%) i/s     (2.08 s/i) -      3.000 in   6.225603s
              ripper      0.645 (± 0.0%) i/s     (1.55 s/i) -      4.000 in   6.205636s

Comparison:
              ripper:        0.6 i/s
               prism:        0.5 i/s - 1.34x  slower
```

`vernier` tells me it does `method_missing` even for explicitly defined methods like `location`.

https://github.com/ruby/prism/commit/2ea81398cc

1 files changed, 22 insertions, 8 deletions

diff --git a/lib/prism/lex_compat.rb b/lib/prism/lex_compat.rb
index 597e63c73e..c14a9f1603 100644
--- a/lib/prism/lex_compat.rb
+++ b/lib/prism/lex_compat.rb
@@ -1,8 +1,6 @@
 # frozen_string_literal: true
 # :markup: markdown
 
-require "delegate"
-
 module Prism
   # This class is responsible for lexing the source using prism and then
   # converting those tokens to be compatible with Ripper. In the vast majority
@@ -201,27 +199,43 @@ module Prism
     # When we produce tokens, we produce the same arrays that Ripper does.
     # However, we add a couple of convenience methods onto them to make them a
     # little easier to work with. We delegate all other methods to the array.
-    class Token < SimpleDelegator
-      # @dynamic initialize, each, []
+    class Token < BasicObject
+      # Create a new token object with the given ripper-compatible array.
+      def initialize(array)
+        @array = array
+      end
 
       # The location of the token in the source.
       def location
-        self[0]
+        @array[0]
       end
 
       # The type of the token.
       def event
-        self[1]
+        @array[1]
       end
 
       # The slice of the source that this token represents.
       def value
-        self[2]
+        @array[2]
       end
 
       # The state of the lexer when this token was produced.
       def state
-        self[3]
+        @array[3]
+      end
+
+      # We want to pretend that this is just an Array.
+      def ==(other) # :nodoc:
+        @array == other
+      end
+
+      def respond_to_missing?(name, include_private = false) # :nodoc:
+        @array.respond_to?(name, include_private)
+      end
+
+      def method_missing(name, ...) # :nodoc:
+        @array.send(name, ...)
       end
     end